Prediction of serum testosterone before and after an exercise program using physiological and personality variables.

This study investigated the relationships between serum testosterone level and selected physiological and personality variables in 58 males (21-61 years) before and after a four month physical fitness program consisting of jogging, calisthenics, and recreational activities. Physical fitness scores were obtained for each subject using a regression equation. Serum testosterone was determined using a radioimmunoassay technique. The multiple correlation was used to determine the relationship between serum testosterone and 19 independent variables consisting of age, height, and weight, and percent lean body weight, systolic and diastolic blood pressures, ten factors from the Cattell 16 PF and the three scales of the Eysenck Personality Inventory. A stepwise linear regression was used to identify the predictive power of each of the independent variables. The subjects improved significantly in physical fitness (p < 0.01). The multiple correlations both at the pre-test (R=0.77; R2=0.59) and post-test (R=0.67; R2=0.46) were significant at the 0.01 level. Neuroticism, factors dealing with emotional stability, aggression, and intelligence were powerful predictors initially. Similar results were obtained at the post-test except for the addition of percent lean as a powerful predictor and the absence of factors dealing with intelligence and aggression. The emergence of percent lean was seen to reflect changes in body composition associated with testosterone secretion. The absence of aggression at the post-test was interpreted in light of free catecholamine excretion associated with habitual exercise.